This Exam P sample reference tests Conditional Probability. The claim weights for the no-, moderate-, and heavy-lifting groups are 0.02, 0.04, and 0.02. Moderate or heavy lifting therefore accounts for 0.06 of the total 0.08 claim probability, giving 0.75, choice A.
These notes identify the calculation error associated with each wrong letter when that error is reproducible.
BThe value 0.81 would leave posterior probability 0.19 for no lifting. The verified no-lifting claim weight gives 0.02/0.08=0.25 instead.
CThe value 0.85 is approximately (0.08+0.20)/(0.05+0.08+0.20). That calculation adds conditional claim rates without weighting them by the group proportions.
DThe value 0.86 would leave posterior probability 0.14 for no lifting, understating the no-lifting joint claim weight.
EThe value 0.89 would leave only 0.11 posterior probability for no lifting. Bayes' normalization shows that the correct complement is 0.25.
Original practice · fully worked
Original variant: infer a source mix from an overall event rate
Service requests are classified as low, medium, or high complexity. Low-complexity requests form half of all requests. The high-complexity proportion is unknown, and the remainder are medium complexity. Escalation rates for the three classes are 2%, 5%, and 20%, respectively, while 5.6% of all requests are escalated. Given an escalation, calculate the probability that the request was high complexity.
A 0.028
B 0.056
C 0.140
D 0.360
E 0.500
Variant answer in brief
If h is the high-complexity proportion, total probability gives 0.035+0.15h=0.056, so h=0.14. The high-complexity escalation weight is 0.14(0.20)=0.028, which is half of all escalations, choice E.
Setup
Setup
Let h be the high-complexity proportion; then the medium-complexity proportion is 0.50-h.
Pr(L)=0.50
Pr(M)=0.50−h
Pr(H)=h
Model
Model
Use the stated overall escalation probability to determine the unknown mixture proportion.
0.50(0.02)+(0.50−h)(0.05)+h(0.20)=0.056
Compute
Compute
Solve for the high-complexity share, then apply Bayes' theorem.
The 3108-page Probability Proof Manual reorganizes 718 verified Exam P solutions by syllabus skill and adds formula proofs, error patterns, and original worked practice.